AI Platform Administrator

Our client, a global company that handles claims administration, loss adjusting, and benefits administration, is actively looking for an AI Platform Administrator to join their team! This role is fully remote in the United States. ***This is a contract to hire position*** Responsibilities Feature, connector, app, and agent governance Maintain the allowlist of approved connectors, apps, plugins, and extensions across all three platforms, and route every new request through security review and the standard IT governance process before enablement. Administer Claude connectors and the organizational Skills catalog, including version control, ownership assignment, and retirement of superseded Skills. Administer ChatGPT Enterprise workspace settings, including plugin and app availability, action and approval controls, custom GPT policy, and Workspace Agents settings — which are disabled by default and require explicit risk review before enablement, particularly where agents would publish using personal or shared connections. Administer Microsoft 365 Copilot controls in the Microsoft 365 admin center, including agent enablement, blocking, and assignment through integrated apps, connector availability, and coordination with the Power Platform team on Copilot Studio publishing and environment policy. Maintain a reviewed inventory of agents, custom GPTs, and Skills in each tenant on a scheduled cadence, so that capabilities introduced through vendor app updates or self-service authoring do not bypass review. Configure sharing and visibility defaults, including Claude project visibility and restriction of public projects, ChatGPT workspace sharing policy, and Copilot agent sharing scope. Level 2 triage and support operations Serve as the designated Level 2 escalation point for all incidents and service requests relating to Claude, ChatGPT, and Copilot routed from the Service Desk (Level 1), owning queue triage, prioritization, and adherence to published service levels. Triage access and authentication failures and partner with IAM to resolve them, including federation and assertion errors, provisioning and group-membership gaps, entitlement and seat-type mismatches, and orphaned or duplicated accounts; own the platform-side diagnosis and escalate the provisioning or identity-side fix to the accountable team. Troubleshoot Claude Code client issues, including enterprise authentication, re-authentication after account or client changes, client version updates, migration from API key–based access to enterprise seats, and terminal and operating-system dependencies across Windows, macOS, and Linux/WSL. Troubleshoot ChatGPT Enterprise desktop, Office add-in, and developer tooling issues, including workspace and seat scoping, plugin and connector authorization, and role-based access to restricted capabilities. Troubleshoot Microsoft 365 Copilot issues, including license and eligibility checks, agent availability, grounding and indexing behavior across SharePoint and Exchange content, add-in availability in Office clients, and permission-inherited results. Governance, compliance, and audit Monitor audit and activity logging across platforms and support the flow of log data into enterprise security monitoring — Claude audit logs, the ChatGPT Enterprise compliance logs platform, and Microsoft Purview audit for Copilot and agent activity. Administer data retention settings in line with Sedgwick records management, Legal, and Privacy requirements, and document the approved retention position for each platform. Support authorized use of vendor compliance APIs and organization data export for data loss prevention, eDiscovery, and regulatory requests, acting only on validated requests from Legal, Privacy, or Information Security. Maintain the evidence set required for third-party risk, security, and privacy reviews, including certification artifacts, data processing agreements, and any business associate agreement obligations, and track vendor control changes that affect Sedgwick’s approved posture. Adoption analytics and reporting Produce recurring adoption reporting from each platform’s native analytics — Claude usage and Claude Code analytics, ChatGPT workspace analytics and usage exports, and Copilot usage reporting and analytics — including weekly active users as a percentage of provisioned seats, activity per user, and feature adoption. Enablement and change management support Support all-staff introductory sessions, department-level enablement built around real workflows, and recurring office hours during initial rollout and for each new team onboarded. Operate the internal support and community channel, maintain the self-service resource hub, and curate a library of approved prompts, Skills, GPTs, and agents with the AI Center of Excellence. Enable and support the champion network with early access, advanced guidance, and demonstration material. Required Skills & Experience Bachelor’s degree in information technology, computer science, or a related field, or equivalent professional experience. 5+ administering enterprise SaaS platforms in a regulated or large-enterprise environment, including at least two years as a named Level 2 escalation owner. Hands-on administration experience in a Microsoft 365 tenant, including the Microsoft 365 admin center, licensing and service plan assignment, and integrated app or add-in governance. Working knowledge of enterprise single sign-on and SCIM provisioning concepts sufficient to triage access failures accurately and partner effectively with IAM, without owning identity provider configuration. Working experience in an ITSM platform (ServiceNow preferred) across incident, request, problem, and knowledge management, operating to defined service levels. Ability to interpret audit logs and administrative analytics, use REST APIs for administrative tasks, and write or adapt basic scripts (PowerShell or Python). Command-line and terminal proficiency sufficient to support developer tooling installation and authentication across Windows, macOS, and Linux/WSL. Strong technical writing skills, with a track record of producing runbooks, knowledge articles, and operating policies that reduce escalation volume.